Advanced Structural Equation Models using Mplus
Dates:
February 20, 2012 - February 24, 2012
Instructor:
Assoc. Prof. Everarda Cunningham Mplus is rapidly becoming the program of choice for the analysis of SEMs. Mplus offers a general modelling framework that allows both the modelling of cross-sectional and longitudinal data using observed variables that are a combination of continuous and categorical variables.
In addition, Mplus allows for multilevel modelling structures. Multilevel modelling takes into account that individuals interact with their social contexts. In other words, individuals are influenced by the social groups or contexts to which they belong (e.g. patients in hospitals; voters in electorates; students in classes in schools).
Standard multivariate models are not appropriate for such data because the assumption of independent observations underlying such analyses is not valid. In instances such as these, which are frequent in social and political research, data are nested in a hierarchical manner and multilevel modelling is the appropriate data analysis technique. Mplus also has the capacity to analyse models involving categorical nominal and ordinal dependent variables (e.g. latent class analysis).
Prerequisites:
This course in an advanced course and therefore it is essential that participants have a basic understanding of Structural Equation Modelling (SEM). In other words, participants must have completed Introduction to Structural Equation Modelling using Amos™, or another course using Amos, Lisrel, EQS, Mplus or an equivalent package. It is also assumed that participants have previous computer experience with statistical data analysis packages such as SPSS.
Learning Objectives:
The course is designed as a comprehensive coverage of applied SEM techniques using the Mplus statistical software package. The first part of the course is focused on using the Mplus program to specify and interpret common models in SEM such as path analytic models, confirmatory factor analysis (CFA) models, and full SEM models involving both continuous and ordinal (e.g., logistic regression models) dependent variables.
The second part of the course introduces multi-group comparisons by initially testing measurement model invariance both cross-sectionally and longitudinally. More complex models including interaction effects for observed and latent constructs and longitudinal models are then presented. The third part of the course focuses on the conceptual and statistical underpinnings of multi-level modelling, including analysing longitudinal growth models, calculation of intra-class correlations and specifying and analysing multilevel models in Mplus.
Course participants will be provided with instruction and hands-on experience in the use of Mplus to estimate parameters implied by the various types of models. Course notes with worked examples and references will be provided. Participants may wish to download the Mplus manual listed below.
Course Text:
Cunningham, E. (2010). A Practical Guide to Structural Equation Modelling using Mplus™. Melbourne: Statsline. These notes will be distributed to all course participants.
References:
Hancock, G. R., & Mueller, R. O. (Eds.) (2006). Structural Equation Modeling: A Second Course. Information Age Publishing, Inc. Chs. 8, 11.
Hox, J. J. (2002). Multilevel Analysis: Techniques and Applications. Lawrence Erlbaum.
Muthén, L.K., & Muthén, B.O. 1998–2007. Mplus User’s Guide. Fifth Edition. Muthén & Muthén. (Downloadable from www.statmodel.com).
